L1.08

Confidence Displays and Their Misreadings

Cards in this group · 5

  1. L1.08.1Confidence is the model’s self-report; the user has no independent way to verify it
  2. L1.08.2Percents are read as frequency promises, and most model numbers are uncalibrated
  3. L1.08.3Bands are less over-read than continuous numbers, at the cost of hiding within-band differences
  4. L1.08.4Showing uniformly high confidence on a whole batch provides no discrimination
  5. L1.08.5Confidence is worth showing only when the user can change the next action because of it